Andrei Zimin
Technical Product Leader · AI Safety · Angel Investor

Hi, I'm Andrei.

I'm an AI-native product manager who likes to bring ideas to life. My focus is Developer Experience and Developer Productivity. These days I work on AI Safety — making AI agents safer to put into production.

On the side, I prototype my own ideas to find out if they're any good. If something's easier to show than to spec, I'll usually just go build it.

A few things I've built

Work

AWS0→1AI SafetyBeta shipped

AI Safety

As AI agents started taking real actions in production, teams had no quick way to tell a risky move from a safe one. I proposed and championed a tool that does it for them — and made one deliberate call along the way: keep it deterministic, trained on past incidents, instead of reaching for a Generative AI model.

Guarding against black boxes shouldn't mean adding another one.

AsanaInfra PlatformObservabilityEnterprise

Observability & infrastructure platform

At Asana I owned observability and infrastructure for the company's own developers — and feature scaling for its largest enterprise customers. I championed consolidating a sprawl of observability tools into one coherent platform, cutting cost and friction while lifting developer satisfaction, and led cross-functional teams to clear the scaling blockers between big customers and the product.

Cut the tool sprawl, and every team moves faster.

VMware0→1DevExBackstage.io

Tanzu Developer Portal

I owned Tanzu's Backstage.io-based internal developer portal and drove a 0→1 product from customer research to launch — growing it from a Kubernetes console into a full developer-experience portal, and authoring the brief that brought generative AI into it. I presented it on stage at VMware Explore.

Meet developers where they already work.

Co-founderRoboticsComputer vision

Friendly Robots Company

Earlier on, I co-founded a robotics startup building computer-vision cleaning robots. We had a founder's constraint — prove the thing works before the runway runs out — and the lesson stuck with me: assemble a scrappy MVP, get something working fast, and let real behavior tell you what to build next.

Get to a prototype fast - how I approach product today.

The short version

About

I'm an AI-native product manager, happiest at the seam between technical architecture and product judgment. Right now that means AI Safety and developer productivity.

I build to think. I've deployed agents on cloud infrastructure, wired up Home Assistant projects around the house, and I'd genuinely rather hand a teammate a working prototype than a ten-page spec. The frameworks I care about — customer obsession, working backwards, falsifiable experiments, reversible decisions — are all just ways of getting to honest signal faster.

The path here ran through enterprise developer platforms and infrastructure. Prior to AWS, I led observability and platform work at Asana, and owned the Tanzu Developer Portal — a Backstage.io-based internal developer portal — at VMware, where I took a 0→1 product from customer research to launch and presented it on stage at VMware Explore. Earlier still: co-founding Friendly Robots, strategy consulting at Dell, and business development for the semiconductor industry.

I'm an economist and statistician by training — a PhD in Mathematical Methods in Economics from Lomonosov Moscow State University and an MBA from Berkeley Haas. Based in the San Francisco Bay Area.

Say hi

Let's chat.

I'm always up for a conversation about product strategy, AI, or robots — whether it's a good idea, a collaboration, or just something I could help you with. LinkedIn's the fastest way to reach me.

Happy to send my résumé — connect on LinkedIn and I'll send it over.